abb-opencode-local-rag 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +21 -0
- package/README.de.md +416 -0
- package/README.es.md +416 -0
- package/README.fr.md +416 -0
- package/README.md +491 -0
- package/README.pt-BR.md +416 -0
- package/README.zh-CN.md +416 -0
- package/dist/bin/install-skills.d.ts +20 -0
- package/dist/bin/install-skills.d.ts.map +1 -0
- package/dist/bin/install-skills.js +195 -0
- package/dist/bin/install-skills.js.map +1 -0
- package/dist/chunker/index.d.ts +24 -0
- package/dist/chunker/index.d.ts.map +1 -0
- package/dist/chunker/index.js +2 -0
- package/dist/chunker/index.js.map +1 -0
- package/dist/chunker/semantic-chunker.d.ts +97 -0
- package/dist/chunker/semantic-chunker.d.ts.map +1 -0
- package/dist/chunker/semantic-chunker.js +294 -0
- package/dist/chunker/semantic-chunker.js.map +1 -0
- package/dist/chunker/sentence-splitter.d.ts +28 -0
- package/dist/chunker/sentence-splitter.d.ts.map +1 -0
- package/dist/chunker/sentence-splitter.js +219 -0
- package/dist/chunker/sentence-splitter.js.map +1 -0
- package/dist/cli/common.d.ts +65 -0
- package/dist/cli/common.d.ts.map +1 -0
- package/dist/cli/common.js +138 -0
- package/dist/cli/common.js.map +1 -0
- package/dist/cli/delete.d.ts +8 -0
- package/dist/cli/delete.d.ts.map +1 -0
- package/dist/cli/delete.js +173 -0
- package/dist/cli/delete.js.map +1 -0
- package/dist/cli/file-collection.d.ts +2 -0
- package/dist/cli/file-collection.d.ts.map +1 -0
- package/dist/cli/file-collection.js +53 -0
- package/dist/cli/file-collection.js.map +1 -0
- package/dist/cli/ingest.d.ts +100 -0
- package/dist/cli/ingest.d.ts.map +1 -0
- package/dist/cli/ingest.js +363 -0
- package/dist/cli/ingest.js.map +1 -0
- package/dist/cli/list.d.ts +35 -0
- package/dist/cli/list.d.ts.map +1 -0
- package/dist/cli/list.js +210 -0
- package/dist/cli/list.js.map +1 -0
- package/dist/cli/options.d.ts +100 -0
- package/dist/cli/options.d.ts.map +1 -0
- package/dist/cli/options.js +241 -0
- package/dist/cli/options.js.map +1 -0
- package/dist/cli/query.d.ts +24 -0
- package/dist/cli/query.d.ts.map +1 -0
- package/dist/cli/query.js +191 -0
- package/dist/cli/query.js.map +1 -0
- package/dist/cli/read-neighbors.d.ts +11 -0
- package/dist/cli/read-neighbors.d.ts.map +1 -0
- package/dist/cli/read-neighbors.js +224 -0
- package/dist/cli/read-neighbors.js.map +1 -0
- package/dist/cli/status.d.ts +8 -0
- package/dist/cli/status.d.ts.map +1 -0
- package/dist/cli/status.js +80 -0
- package/dist/cli/status.js.map +1 -0
- package/dist/cli/sync.d.ts +8 -0
- package/dist/cli/sync.d.ts.map +1 -0
- package/dist/cli/sync.js +244 -0
- package/dist/cli/sync.js.map +1 -0
- package/dist/cli-main.d.ts +12 -0
- package/dist/cli-main.d.ts.map +1 -0
- package/dist/cli-main.js +63 -0
- package/dist/cli-main.js.map +1 -0
- package/dist/embedder/index.d.ts +85 -0
- package/dist/embedder/index.d.ts.map +1 -0
- package/dist/embedder/index.js +284 -0
- package/dist/embedder/index.js.map +1 -0
- package/dist/features/list.d.ts +37 -0
- package/dist/features/list.d.ts.map +1 -0
- package/dist/features/list.js +40 -0
- package/dist/features/list.js.map +1 -0
- package/dist/features/sync.d.ts +207 -0
- package/dist/features/sync.d.ts.map +1 -0
- package/dist/features/sync.js +380 -0
- package/dist/features/sync.js.map +1 -0
- package/dist/index.d.ts +3 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +53 -0
- package/dist/index.js.map +1 -0
- package/dist/ingest/compute.d.ts +86 -0
- package/dist/ingest/compute.d.ts.map +1 -0
- package/dist/ingest/compute.js +177 -0
- package/dist/ingest/compute.js.map +1 -0
- package/dist/ingest/file.d.ts +27 -0
- package/dist/ingest/file.d.ts.map +1 -0
- package/dist/ingest/file.js +67 -0
- package/dist/ingest/file.js.map +1 -0
- package/dist/ingest/visual.d.ts +45 -0
- package/dist/ingest/visual.d.ts.map +1 -0
- package/dist/ingest/visual.js +234 -0
- package/dist/ingest/visual.js.map +1 -0
- package/dist/parser/docx-parser.d.ts +12 -0
- package/dist/parser/docx-parser.d.ts.map +1 -0
- package/dist/parser/docx-parser.js +328 -0
- package/dist/parser/docx-parser.js.map +1 -0
- package/dist/parser/html-parser.d.ts +18 -0
- package/dist/parser/html-parser.d.ts.map +1 -0
- package/dist/parser/html-parser.js +102 -0
- package/dist/parser/html-parser.js.map +1 -0
- package/dist/parser/index.d.ts +214 -0
- package/dist/parser/index.d.ts.map +1 -0
- package/dist/parser/index.js +454 -0
- package/dist/parser/index.js.map +1 -0
- package/dist/parser/pdf-extract.d.ts +81 -0
- package/dist/parser/pdf-extract.d.ts.map +1 -0
- package/dist/parser/pdf-extract.js +112 -0
- package/dist/parser/pdf-extract.js.map +1 -0
- package/dist/parser/pdf-filter.d.ts +117 -0
- package/dist/parser/pdf-filter.d.ts.map +1 -0
- package/dist/parser/pdf-filter.js +528 -0
- package/dist/parser/pdf-filter.js.map +1 -0
- package/dist/parser/title-extractor.d.ts +69 -0
- package/dist/parser/title-extractor.d.ts.map +1 -0
- package/dist/parser/title-extractor.js +145 -0
- package/dist/parser/title-extractor.js.map +1 -0
- package/dist/pdf-visual/captioner.d.ts +16 -0
- package/dist/pdf-visual/captioner.d.ts.map +1 -0
- package/dist/pdf-visual/captioner.js +63 -0
- package/dist/pdf-visual/captioner.js.map +1 -0
- package/dist/pdf-visual/captioners/fast.d.ts +7 -0
- package/dist/pdf-visual/captioners/fast.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/fast.js +103 -0
- package/dist/pdf-visual/captioners/fast.js.map +1 -0
- package/dist/pdf-visual/captioners/quality.d.ts +7 -0
- package/dist/pdf-visual/captioners/quality.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/quality.js +127 -0
- package/dist/pdf-visual/captioners/quality.js.map +1 -0
- package/dist/pdf-visual/captioners/shared.d.ts +44 -0
- package/dist/pdf-visual/captioners/shared.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/shared.js +104 -0
- package/dist/pdf-visual/captioners/shared.js.map +1 -0
- package/dist/pdf-visual/detector.d.ts +9 -0
- package/dist/pdf-visual/detector.d.ts.map +1 -0
- package/dist/pdf-visual/detector.js +234 -0
- package/dist/pdf-visual/detector.js.map +1 -0
- package/dist/pdf-visual/index.d.ts +13 -0
- package/dist/pdf-visual/index.d.ts.map +1 -0
- package/dist/pdf-visual/index.js +45 -0
- package/dist/pdf-visual/index.js.map +1 -0
- package/dist/pdf-visual/renderer.d.ts +9 -0
- package/dist/pdf-visual/renderer.d.ts.map +1 -0
- package/dist/pdf-visual/renderer.js +177 -0
- package/dist/pdf-visual/renderer.js.map +1 -0
- package/dist/pdf-visual/types.d.ts +62 -0
- package/dist/pdf-visual/types.d.ts.map +1 -0
- package/dist/pdf-visual/types.js +32 -0
- package/dist/pdf-visual/types.js.map +1 -0
- package/dist/server/error-utils.d.ts +79 -0
- package/dist/server/error-utils.d.ts.map +1 -0
- package/dist/server/error-utils.js +148 -0
- package/dist/server/error-utils.js.map +1 -0
- package/dist/server/index.d.ts +258 -0
- package/dist/server/index.d.ts.map +1 -0
- package/dist/server/index.js +1104 -0
- package/dist/server/index.js.map +1 -0
- package/dist/server/list-scanner.d.ts +52 -0
- package/dist/server/list-scanner.d.ts.map +1 -0
- package/dist/server/list-scanner.js +72 -0
- package/dist/server/list-scanner.js.map +1 -0
- package/dist/server/tool-definitions.d.ts +8 -0
- package/dist/server/tool-definitions.d.ts.map +1 -0
- package/dist/server/tool-definitions.js +181 -0
- package/dist/server/tool-definitions.js.map +1 -0
- package/dist/server/tool-input.d.ts +37 -0
- package/dist/server/tool-input.d.ts.map +1 -0
- package/dist/server/tool-input.js +216 -0
- package/dist/server/tool-input.js.map +1 -0
- package/dist/server/types.d.ts +331 -0
- package/dist/server/types.d.ts.map +1 -0
- package/dist/server/types.js +3 -0
- package/dist/server/types.js.map +1 -0
- package/dist/server-main.d.ts +46 -0
- package/dist/server-main.d.ts.map +1 -0
- package/dist/server-main.js +242 -0
- package/dist/server-main.js.map +1 -0
- package/dist/utils/base-dirs.d.ts +212 -0
- package/dist/utils/base-dirs.d.ts.map +1 -0
- package/dist/utils/base-dirs.js +422 -0
- package/dist/utils/base-dirs.js.map +1 -0
- package/dist/utils/errors.d.ts +24 -0
- package/dist/utils/errors.d.ts.map +1 -0
- package/dist/utils/errors.js +53 -0
- package/dist/utils/errors.js.map +1 -0
- package/dist/utils/limits.d.ts +26 -0
- package/dist/utils/limits.d.ts.map +1 -0
- package/dist/utils/limits.js +28 -0
- package/dist/utils/limits.js.map +1 -0
- package/dist/utils/list-sources.d.ts +47 -0
- package/dist/utils/list-sources.d.ts.map +1 -0
- package/dist/utils/list-sources.js +50 -0
- package/dist/utils/list-sources.js.map +1 -0
- package/dist/utils/raw-data-utils.d.ts +131 -0
- package/dist/utils/raw-data-utils.d.ts.map +1 -0
- package/dist/utils/raw-data-utils.js +255 -0
- package/dist/utils/raw-data-utils.js.map +1 -0
- package/dist/utils/scan.d.ts +126 -0
- package/dist/utils/scan.d.ts.map +1 -0
- package/dist/utils/scan.js +221 -0
- package/dist/utils/scan.js.map +1 -0
- package/dist/utils/scope-match.d.ts +43 -0
- package/dist/utils/scope-match.d.ts.map +1 -0
- package/dist/utils/scope-match.js +87 -0
- package/dist/utils/scope-match.js.map +1 -0
- package/dist/utils/sensitive-path.d.ts +23 -0
- package/dist/utils/sensitive-path.d.ts.map +1 -0
- package/dist/utils/sensitive-path.js +91 -0
- package/dist/utils/sensitive-path.js.map +1 -0
- package/dist/utils/sync-path-key.d.ts +20 -0
- package/dist/utils/sync-path-key.d.ts.map +1 -0
- package/dist/utils/sync-path-key.js +33 -0
- package/dist/utils/sync-path-key.js.map +1 -0
- package/dist/vectordb/index.d.ts +168 -0
- package/dist/vectordb/index.d.ts.map +1 -0
- package/dist/vectordb/index.js +619 -0
- package/dist/vectordb/index.js.map +1 -0
- package/dist/vectordb/search-filters.d.ts +39 -0
- package/dist/vectordb/search-filters.d.ts.map +1 -0
- package/dist/vectordb/search-filters.js +136 -0
- package/dist/vectordb/search-filters.js.map +1 -0
- package/dist/vectordb/types.d.ts +196 -0
- package/dist/vectordb/types.d.ts.map +1 -0
- package/dist/vectordb/types.js +224 -0
- package/dist/vectordb/types.js.map +1 -0
- package/package.json +105 -0
- package/skills/mcp-local-rag/SKILL.md +308 -0
- package/skills/mcp-local-rag/references/cli-reference.md +175 -0
- package/skills/mcp-local-rag/references/html-ingestion.md +78 -0
- package/skills/mcp-local-rag/references/query-optimization.md +57 -0
- package/skills/mcp-local-rag/references/result-refinement.md +56 -0
|
@@ -0,0 +1,234 @@
|
|
|
1
|
+
import { buildChunksAndEmbeddings, createVisualAttachment, findNearestChunk } from './compute.js';
|
|
2
|
+
function horizontalOverlap(left, right) {
|
|
3
|
+
return Math.min(left[2] ?? 0, right[2] ?? 0) >= Math.max(left[0] ?? 0, right[0] ?? 0);
|
|
4
|
+
}
|
|
5
|
+
function xCenter(bbox) {
|
|
6
|
+
return ((bbox[0] ?? 0) + (bbox[2] ?? 0)) / 2;
|
|
7
|
+
}
|
|
8
|
+
function yCenter(bbox) {
|
|
9
|
+
return ((bbox[1] ?? 0) + (bbox[3] ?? 0)) / 2;
|
|
10
|
+
}
|
|
11
|
+
function findInsertionIndex(fragments, region) {
|
|
12
|
+
if (fragments.length === 0)
|
|
13
|
+
return 0;
|
|
14
|
+
let candidates = fragments
|
|
15
|
+
.map((fragment, index) => ({ fragment, index }))
|
|
16
|
+
.filter(({ fragment }) => horizontalOverlap(fragment.bbox, region.bbox));
|
|
17
|
+
if (candidates.length === 0) {
|
|
18
|
+
const regionCenter = xCenter(region.bbox);
|
|
19
|
+
const minimumDistance = Math.min(...fragments.map((fragment) => Math.abs(xCenter(fragment.bbox) - regionCenter)));
|
|
20
|
+
candidates = fragments
|
|
21
|
+
.map((fragment, index) => ({ fragment, index }))
|
|
22
|
+
.filter(({ fragment }) => Math.abs(xCenter(fragment.bbox) - regionCenter) === minimumDistance);
|
|
23
|
+
}
|
|
24
|
+
const regionMidpoint = yCenter(region.bbox);
|
|
25
|
+
const preceding = candidates.filter(({ fragment }) => yCenter(fragment.bbox) <= regionMidpoint);
|
|
26
|
+
if (preceding.length > 0) {
|
|
27
|
+
return Math.max(...preceding.map(({ index }) => index + 1));
|
|
28
|
+
}
|
|
29
|
+
return Math.min(...candidates.map(({ index }) => index));
|
|
30
|
+
}
|
|
31
|
+
function fallbackFragments(page) {
|
|
32
|
+
if (page.textFragments.length > 0 || page.text.length === 0)
|
|
33
|
+
return page.textFragments;
|
|
34
|
+
return [
|
|
35
|
+
{
|
|
36
|
+
pageNum: page.pageNum,
|
|
37
|
+
blockOrdinal: 0,
|
|
38
|
+
lineOrdinal: 0,
|
|
39
|
+
fragmentOrdinal: 0,
|
|
40
|
+
bbox: [0, 0, 0, 0],
|
|
41
|
+
text: page.text,
|
|
42
|
+
pageTextStart: 0,
|
|
43
|
+
pageTextEnd: page.text.length,
|
|
44
|
+
},
|
|
45
|
+
];
|
|
46
|
+
}
|
|
47
|
+
export function buildOrderedVisualDocument(pages, processedRegions) {
|
|
48
|
+
let text = '';
|
|
49
|
+
let visualIndex = 0;
|
|
50
|
+
const atomicRanges = [];
|
|
51
|
+
const regions = [];
|
|
52
|
+
const consumedRegions = new Set();
|
|
53
|
+
for (const page of pages) {
|
|
54
|
+
const pageFragments = fallbackFragments(page);
|
|
55
|
+
const positioned = processedRegions
|
|
56
|
+
.filter((region) => region.pageNum === page.pageNum)
|
|
57
|
+
.map((region) => ({
|
|
58
|
+
region,
|
|
59
|
+
insertionIndex: findInsertionIndex(pageFragments, region),
|
|
60
|
+
}))
|
|
61
|
+
.sort((left, right) => left.insertionIndex - right.insertionIndex ||
|
|
62
|
+
left.region.detectionIndex - right.region.detectionIndex);
|
|
63
|
+
const pageVisualIndices = new Map(positioned.map((item, index) => [item.region, visualIndex + index]));
|
|
64
|
+
const insertionOffsets = new Map();
|
|
65
|
+
for (const item of positioned) {
|
|
66
|
+
const preceding = pageFragments[item.insertionIndex - 1];
|
|
67
|
+
const following = pageFragments[item.insertionIndex];
|
|
68
|
+
insertionOffsets.set(item.region, Math.max(0, Math.min(page.text.length, preceding?.pageTextEnd ?? following?.pageTextStart ?? 0)));
|
|
69
|
+
}
|
|
70
|
+
const insertionGroups = new Map();
|
|
71
|
+
for (const item of positioned) {
|
|
72
|
+
const offset = insertionOffsets.get(item.region) ?? 0;
|
|
73
|
+
const group = insertionGroups.get(offset) ?? [];
|
|
74
|
+
group.push(item.region);
|
|
75
|
+
insertionGroups.set(offset, group);
|
|
76
|
+
}
|
|
77
|
+
let pageText = '';
|
|
78
|
+
let pageCursor = 0;
|
|
79
|
+
const captionRanges = new Map();
|
|
80
|
+
const anchorOffsets = new Map();
|
|
81
|
+
for (const [offset, group] of [...insertionGroups].sort((left, right) => left[0] - right[0])) {
|
|
82
|
+
pageText += page.text.slice(pageCursor, offset);
|
|
83
|
+
const hasCaption = group.some((region) => region.caption !== null);
|
|
84
|
+
if (hasCaption) {
|
|
85
|
+
const trailingNewlines = pageText.match(/\n*$/)?.[0].length ?? 0;
|
|
86
|
+
pageText += pageText.length > 0 ? '\n'.repeat(Math.max(0, 2 - trailingNewlines)) : '';
|
|
87
|
+
}
|
|
88
|
+
let captionIndex = 0;
|
|
89
|
+
for (const region of group) {
|
|
90
|
+
if (region.caption === null) {
|
|
91
|
+
anchorOffsets.set(region, pageText.length);
|
|
92
|
+
continue;
|
|
93
|
+
}
|
|
94
|
+
if (captionIndex > 0)
|
|
95
|
+
pageText += '\n\n';
|
|
96
|
+
const start = pageText.length;
|
|
97
|
+
pageText += `[Visual content on page ${page.pageNum}, visual ${pageVisualIndices.get(region)}: ${region.caption}]`;
|
|
98
|
+
captionRanges.set(region, { start, end: pageText.length });
|
|
99
|
+
anchorOffsets.set(region, start);
|
|
100
|
+
captionIndex += 1;
|
|
101
|
+
}
|
|
102
|
+
if (hasCaption && offset < page.text.length) {
|
|
103
|
+
const leadingNewlines = page.text.slice(offset).match(/^\n*/)?.[0].length ?? 0;
|
|
104
|
+
pageText += '\n'.repeat(Math.max(0, 2 - leadingNewlines));
|
|
105
|
+
}
|
|
106
|
+
pageCursor = offset;
|
|
107
|
+
}
|
|
108
|
+
pageText += page.text.slice(pageCursor);
|
|
109
|
+
const pageSeparator = text.length > 0 && pageText.length > 0 ? '\n\n' : '';
|
|
110
|
+
const pageStart = text.length + pageSeparator.length;
|
|
111
|
+
text += pageSeparator + pageText;
|
|
112
|
+
for (const positionedRegion of positioned) {
|
|
113
|
+
const currentVisualIndex = visualIndex++;
|
|
114
|
+
const relativeCaptionRange = captionRanges.get(positionedRegion.region);
|
|
115
|
+
const captionRange = relativeCaptionRange
|
|
116
|
+
? {
|
|
117
|
+
start: pageStart + relativeCaptionRange.start,
|
|
118
|
+
end: pageStart + relativeCaptionRange.end,
|
|
119
|
+
}
|
|
120
|
+
: undefined;
|
|
121
|
+
if (captionRange)
|
|
122
|
+
atomicRanges.push(captionRange);
|
|
123
|
+
const anchorOffset = anchorOffsets.get(positionedRegion.region);
|
|
124
|
+
if (anchorOffset === undefined)
|
|
125
|
+
throw new Error('Visual region has no anchor offset');
|
|
126
|
+
regions.push({
|
|
127
|
+
...positionedRegion.region,
|
|
128
|
+
visualIndex: currentVisualIndex,
|
|
129
|
+
anchorOffset: pageStart + anchorOffset,
|
|
130
|
+
...(captionRange ? { captionRange } : {}),
|
|
131
|
+
});
|
|
132
|
+
consumedRegions.add(positionedRegion.region);
|
|
133
|
+
}
|
|
134
|
+
}
|
|
135
|
+
if (consumedRegions.size !== processedRegions.length) {
|
|
136
|
+
throw new Error('A processed visual region has no matching PDF page');
|
|
137
|
+
}
|
|
138
|
+
return { text, atomicRanges, regions };
|
|
139
|
+
}
|
|
140
|
+
function assignVisualAttachments(document, chunks, attachments) {
|
|
141
|
+
const attachmentsByChunkIndex = new Map();
|
|
142
|
+
if (chunks.length === 0)
|
|
143
|
+
return attachmentsByChunkIndex;
|
|
144
|
+
const regionByVisualIndex = new Map(document.regions.map((region) => [region.visualIndex, region]));
|
|
145
|
+
const seenVisualIndices = new Set();
|
|
146
|
+
for (const attachment of [...attachments].sort((left, right) => left.imageIndex - right.imageIndex)) {
|
|
147
|
+
if (seenVisualIndices.has(attachment.imageIndex)) {
|
|
148
|
+
throw new Error(`Duplicate visual attachment index: ${attachment.imageIndex}`);
|
|
149
|
+
}
|
|
150
|
+
seenVisualIndices.add(attachment.imageIndex);
|
|
151
|
+
const region = regionByVisualIndex.get(attachment.imageIndex);
|
|
152
|
+
if (!region)
|
|
153
|
+
throw new Error(`Visual attachment has no ordered region: ${attachment.imageIndex}`);
|
|
154
|
+
const owner = findNearestChunk(chunks, region.anchorOffset);
|
|
155
|
+
if (!owner)
|
|
156
|
+
throw new Error(`Visual ${attachment.imageIndex} has no owning chunk`);
|
|
157
|
+
const current = attachmentsByChunkIndex.get(owner.index) ?? [];
|
|
158
|
+
current.push(attachment);
|
|
159
|
+
attachmentsByChunkIndex.set(owner.index, current);
|
|
160
|
+
}
|
|
161
|
+
return attachmentsByChunkIndex;
|
|
162
|
+
}
|
|
163
|
+
async function processImageOnlyRegions(regions, doc) {
|
|
164
|
+
const renderer = await import('../pdf-visual/renderer.js');
|
|
165
|
+
const processed = [];
|
|
166
|
+
let omittedImageCount = 0;
|
|
167
|
+
for (const region of regions) {
|
|
168
|
+
try {
|
|
169
|
+
const image = await renderer.renderPdfRendition(doc, region.pageNum, region.bbox, region.evidence);
|
|
170
|
+
processed.push({ ...region, caption: null, rendition: image });
|
|
171
|
+
}
|
|
172
|
+
catch {
|
|
173
|
+
omittedImageCount += 1;
|
|
174
|
+
processed.push({ ...region, caption: null });
|
|
175
|
+
}
|
|
176
|
+
}
|
|
177
|
+
return { processed, omittedImageCount };
|
|
178
|
+
}
|
|
179
|
+
export async function prepareVisualPdfChunks(filePath, parser, chunker, embedder, options) {
|
|
180
|
+
const captionerConfig = options.captioner;
|
|
181
|
+
const { doc, title, pages } = await parser.parsePdfPages(filePath, embedder);
|
|
182
|
+
try {
|
|
183
|
+
let processed;
|
|
184
|
+
let omittedImageCount = 0;
|
|
185
|
+
if (captionerConfig !== undefined) {
|
|
186
|
+
const pdfVisual = await import('../pdf-visual/index.js');
|
|
187
|
+
const regions = pdfVisual.detectVisualRegions(pages.map((page) => ({ pageNum: page.pageNum, stextJson: page.stextJson })), doc);
|
|
188
|
+
const captioner = pdfVisual.createCaptioner(captionerConfig);
|
|
189
|
+
processed = await pdfVisual.processVisualRegions(regions, doc, { captioner, includeImages: options.images });
|
|
190
|
+
}
|
|
191
|
+
else {
|
|
192
|
+
const detector = await import('../pdf-visual/detector.js');
|
|
193
|
+
const regions = detector.detectVisualRegions(pages.map((page) => ({ pageNum: page.pageNum, stextJson: page.stextJson })), doc);
|
|
194
|
+
const imageOnly = await processImageOnlyRegions(regions, doc);
|
|
195
|
+
processed = imageOnly.processed;
|
|
196
|
+
omittedImageCount += imageOnly.omittedImageCount;
|
|
197
|
+
}
|
|
198
|
+
const ordered = buildOrderedVisualDocument(pages, processed);
|
|
199
|
+
const { chunks, embeddings } = await buildChunksAndEmbeddings(ordered.text, chunker, embedder, ordered.atomicRanges);
|
|
200
|
+
const attachments = [];
|
|
201
|
+
if (options.images) {
|
|
202
|
+
for (const orderedRegion of ordered.regions) {
|
|
203
|
+
if (!orderedRegion.rendition)
|
|
204
|
+
continue;
|
|
205
|
+
try {
|
|
206
|
+
attachments.push(createVisualAttachment(orderedRegion.visualIndex, orderedRegion.rendition));
|
|
207
|
+
}
|
|
208
|
+
catch {
|
|
209
|
+
omittedImageCount += 1;
|
|
210
|
+
}
|
|
211
|
+
}
|
|
212
|
+
}
|
|
213
|
+
const visualAttachments = assignVisualAttachments(ordered, chunks, attachments);
|
|
214
|
+
return {
|
|
215
|
+
chunks,
|
|
216
|
+
embeddings,
|
|
217
|
+
title,
|
|
218
|
+
text: ordered.text,
|
|
219
|
+
atomicRanges: ordered.atomicRanges,
|
|
220
|
+
visualAttachments,
|
|
221
|
+
omittedImageCount,
|
|
222
|
+
};
|
|
223
|
+
}
|
|
224
|
+
finally {
|
|
225
|
+
try {
|
|
226
|
+
doc.destroy();
|
|
227
|
+
}
|
|
228
|
+
catch (destroyError) {
|
|
229
|
+
const message = destroyError instanceof Error ? destroyError.message : String(destroyError);
|
|
230
|
+
console.warn(`prepareVisualPdfChunks: doc.destroy() failed: ${message}`);
|
|
231
|
+
}
|
|
232
|
+
}
|
|
233
|
+
}
|
|
234
|
+
//# sourceMappingURL=visual.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"visual.js","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAUA,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAA;AA6CjG,SAAS,iBAAiB,CAAC,IAAuB,EAAE,KAAwB;IAC1E,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACvF,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,kBAAkB,CACzB,SAA0C,EAC1C,MAA4B;IAE5B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAA;IACpC,IAAI,UAAU,GAAG,SAAS;SACvB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;SAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,iBAAiB,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAA;IAE1E,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,MAAM,YAAY,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACzC,MAAM,eAAe,GAAG,IAAI,CAAC,GAAG,CAC9B,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC,CAChF,CAAA;QACD,UAAU,GAAG,SAAS;aACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;aAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,KAAK,eAAe,CAAC,CAAA;IAClG,CAAC;IAED,MAAM,cAAc,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,CAAA;IAC/F,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAA;IAC7D,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAA;AAC1D,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAuB;IAChD,IAAI,IAAI,CAAC,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC,aAAa,CAAA;IACtF,OAAO;QACL;YACE,OAAO,EAAE,IAAI,CAAC,OAAO;YACrB,YAAY,EAAE,CAAC;YACf,WAAW,EAAE,CAAC;YACd,eAAe,EAAE,CAAC;YAClB,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YAClB,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,aAAa,EAAE,CAAC;YAChB,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM;SAC9B;KACF,CAAA;AACH,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,KAAmC,EACnC,gBAAkD;IAElD,IAAI,IAAI,GAAG,EAAE,CAAA;IACb,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,OAAO,GAA0B,EAAE,CAAA;IACzC,MAAM,eAAe,GAAG,IAAI,GAAG,EAAyB,CAAA;IAExD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC7C,MAAM,UAAU,GAAG,gBAAgB;aAChC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,CAAC;aACnD,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;YAChB,MAAM;YACN,cAAc,EAAE,kBAAkB,CAAC,aAAa,EAAE,MAAM,CAAC;SAC1D,CAAC,CAAC;aACF,IAAI,CACH,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,IAAI,CAAC,cAAc,GAAG,KAAK,CAAC,cAAc;YAC1C,IAAI,CAAC,MAAM,CAAC,cAAc,GAAG,KAAK,CAAC,MAAM,CAAC,cAAc,CAC3D,CAAA;QACH,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAC/B,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,WAAW,GAAG,KAAK,CAAU,CAAC,CAC7E,CAAA;QAED,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAiC,CAAA;QACjE,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,GAAG,CAAC,CAAC,CAAA;YACxD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;YACpD,gBAAgB,CAAC,GAAG,CAClB,IAAI,CAAC,MAAM,EACX,IAAI,CAAC,GAAG,CACN,CAAC,EACD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,WAAW,IAAI,SAAS,EAAE,aAAa,IAAI,CAAC,CAAC,CACpF,CACF,CAAA;QACH,CAAC;QAED,MAAM,eAAe,GAAG,IAAI,GAAG,EAAmC,CAAA;QAClE,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,MAAM,GAAG,gBAAgB,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACrD,MAAM,KAAK,GAAG,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;YAC/C,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACvB,eAAe,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;QACpC,CAAC;QAED,IAAI,QAAQ,GAAG,EAAE,CAAA;QACjB,IAAI,UAAU,GAAG,CAAC,CAAA;QAClB,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0C,CAAA;QACvE,MAAM,aAAa,GAAG,IAAI,GAAG,EAAiC,CAAA;QAC9D,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YAC7F,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;YAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;YAClE,IAAI,UAAU,EAAE,CAAC;gBACf,MAAM,gBAAgB,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAA;gBAChE,QAAQ,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,gBAAgB,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;YACvF,CAAC;YACD,IAAI,YAAY,GAAG,CAAC,CAAA;YACpB,KAAK,MAAM,MAAM,IAAI,KAAK,EAAE,CAAC;gBAC3B,IAAI,MAAM,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;oBAC5B,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAA;oBAC1C,SAAQ;gBACV,CAAC;gBACD,IAAI,YAAY,GAAG,CAAC;oBAAE,QAAQ,IAAI,MAAM,CAAA;gBACxC,MAAM,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAA;gBAC7B,QAAQ,IAAI,2BAA2B,IAAI,CAAC,OAAO,YAAY,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,MAAM,CAAC,OAAO,GAAG,CAAA;gBAClH,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC1D,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;gBAChC,YAAY,IAAI,CAAC,CAAA;YACnB,CAAC;YACD,IAAI,UAAU,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC5C,MAAM,eAAe,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAA;gBAC9E,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,eAAe,CAAC,CAAC,CAAA;YAC3D,CAAC;YACD,UAAU,GAAG,MAAM,CAAA;QACrB,CAAC;QACD,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAA;QAEvC,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAA;QAC1E,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,CAAA;QACpD,IAAI,IAAI,aAAa,GAAG,QAAQ,CAAA;QAEhC,KAAK,MAAM,gBAAgB,IAAI,UAAU,EAAE,CAAC;YAC1C,MAAM,kBAAkB,GAAG,WAAW,EAAE,CAAA;YACxC,MAAM,oBAAoB,GAAG,aAAa,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;YACvE,MAAM,YAAY,GAAG,oBAAoB;gBACvC,CAAC,CAAC;oBACE,KAAK,EAAE,SAAS,GAAG,oBAAoB,CAAC,KAAK;oBAC7C,GAAG,EAAE,SAAS,GAAG,oBAAoB,CAAC,GAAG;iBAC1C;gBACH,CAAC,CAAC,SAAS,CAAA;YACb,IAAI,YAAY;gBAAE,YAAY,CAAC,IAAI,CAAC,YAAY,CAAC,CAAA;YACjD,MAAM,YAAY,GAAG,aAAa,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;YAC/D,IAAI,YAAY,KAAK,SAAS;gBAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAA;YACrF,OAAO,CAAC,IAAI,CAAC;gBACX,GAAG,gBAAgB,CAAC,MAAM;gBAC1B,WAAW,EAAE,kBAAkB;gBAC/B,YAAY,EAAE,SAAS,GAAG,YAAY;gBACtC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aAC1C,CAAC,CAAA;YACF,eAAe,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IAED,IAAI,eAAe,CAAC,IAAI,KAAK,gBAAgB,CAAC,MAAM,EAAE,CAAC;QACrD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAA;IACvE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,CAAA;AACxC,CAAC;AAED,SAAS,uBAAuB,CAC9B,QAA+B,EAC/B,MAA4B,EAC5B,WAAwC;IAExC,MAAM,uBAAuB,GAAG,IAAI,GAAG,EAA8B,CAAA;IACrE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,uBAAuB,CAAA;IACvD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CACjC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAU,CAAC,CACxE,CAAA;IACD,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAU,CAAA;IAE3C,KAAK,MAAM,UAAU,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,CAC5C,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CACpD,EAAE,CAAC;QACF,IAAI,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YACjD,MAAM,IAAI,KAAK,CAAC,sCAAsC,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QAChF,CAAC;QACD,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC5C,MAAM,MAAM,GAAG,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC7D,IAAI,CAAC,MAAM;YACT,MAAM,IAAI,KAAK,CAAC,4CAA4C,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QAEtF,MAAM,KAAK,GAAG,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,YAAY,CAAC,CAAA;QAC3D,IAAI,CAAC,KAAK;YAAE,MAAM,IAAI,KAAK,CAAC,UAAU,UAAU,CAAC,UAAU,sBAAsB,CAAC,CAAA;QAClF,MAAM,OAAO,GAAG,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;QAC9D,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC,CAAA;IACnD,CAAC;IACD,OAAO,uBAAuB,CAAA;AAChC,CAAC;AAED,KAAK,UAAU,uBAAuB,CACpC,OAA+B,EAC/B,GAAiE;IAKjE,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;IAC1D,MAAM,SAAS,GAA4B,EAAE,CAAA;IAC7C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,kBAAkB,CAC7C,GAAwD,EACxD,MAAM,CAAC,OAAO,EACd,MAAM,CAAC,IAAI,EACX,MAAM,CAAC,QAAQ,CAChB,CAAA;YACD,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAA;QAChE,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;YACtB,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IACD,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAA;AACzC,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,QAAgB,EAChB,MAAuB,EACvB,OAAwB,EACxB,QAA2B,EAC3B,OAAsC;IAEtC,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,CAAA;IACzC,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAA;IAC5E,IAAI,CAAC;QACH,IAAI,SAAkC,CAAA;QACtC,IAAI,iBAAiB,GAAG,CAAC,CAAA;QAEzB,IAAI,eAAe,KAAK,SAAS,EAAE,CAAC;YAClC,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC,CAAA;YACxD,MAAM,OAAO,GAAG,SAAS,CAAC,mBAAmB,CAC3C,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,EAC3E,GAA0D,CAC3D,CAAA;YACD,MAAM,SAAS,GAAG,SAAS,CAAC,eAAe,CAAC,eAAe,CAAC,CAAA;YAC5D,SAAS,GAAG,MAAM,SAAS,CAAC,oBAAoB,CAC9C,OAAO,EACP,GAA2D,EAC3D,EAAE,SAAS,EAAE,aAAa,EAAE,OAAO,CAAC,MAAM,EAAE,CAC7C,CAAA;QACH,CAAC;aAAM,CAAC;YACN,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;YAC1D,MAAM,OAAO,GAAG,QAAQ,CAAC,mBAAmB,CAC1C,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,EAC3E,GAAyD,CAC1D,CAAA;YACD,MAAM,SAAS,GAAG,MAAM,uBAAuB,CAAC,OAAO,EAAE,GAAG,CAAC,CAAA;YAC7D,SAAS,GAAG,SAAS,CAAC,SAAS,CAAA;YAC/B,iBAAiB,IAAI,SAAS,CAAC,iBAAiB,CAAA;QAClD,CAAC;QAED,MAAM,OAAO,GAAG,0BAA0B,CAAC,KAAK,EAAE,SAAS,CAAC,CAAA;QAC5D,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,wBAAwB,CAC3D,OAAO,CAAC,IAAI,EACZ,OAAO,EACP,QAAQ,EACR,OAAO,CAAC,YAAY,CACrB,CAAA;QACD,MAAM,WAAW,GAAuB,EAAE,CAAA;QAC1C,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;YACnB,KAAK,MAAM,aAAa,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;gBAC5C,IAAI,CAAC,aAAa,CAAC,SAAS;oBAAE,SAAQ;gBACtC,IAAI,CAAC;oBACH,WAAW,CAAC,IAAI,CACd,sBAAsB,CAAC,aAAa,CAAC,WAAW,EAAE,aAAa,CAAC,SAAS,CAAC,CAC3E,CAAA;gBACH,CAAC;gBAAC,MAAM,CAAC;oBACP,iBAAiB,IAAI,CAAC,CAAA;gBACxB,CAAC;YACH,CAAC;QACH,CAAC;QACD,MAAM,iBAAiB,GAAG,uBAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAA;QAC/E,OAAO;YACL,MAAM;YACN,UAAU;YACV,KAAK;YACL,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,iBAAiB;YACjB,iBAAiB;SAClB,CAAA;IACH,CAAC;YAAS,CAAC;QACT,IAAI,CAAC;YACH,GAAG,CAAC,OAAO,EAAE,CAAA;QACf,CAAC;QAAC,OAAO,YAAY,EAAE,CAAC;YACtB,MAAM,OAAO,GAAG,YAAY,YAAY,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YAC3F,OAAO,CAAC,IAAI,CAAC,iDAAiD,OAAO,EAAE,CAAC,CAAA;QAC1E,CAAC;IACH,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { AtomicTextRange } from '../chunker/index.js';
|
|
2
|
+
export interface DocxBodyResult {
|
|
3
|
+
content: string;
|
|
4
|
+
atomicRanges: readonly AtomicTextRange[];
|
|
5
|
+
imageAnchors?: readonly {
|
|
6
|
+
offset: number;
|
|
7
|
+
imageIndex: number;
|
|
8
|
+
}[];
|
|
9
|
+
}
|
|
10
|
+
export declare function extractDocxCoreTitle(buffer: Buffer): Promise<string | undefined>;
|
|
11
|
+
export declare function convertDocxDocumentToText(document: Document): DocxBodyResult;
|
|
12
|
+
//# sourceMappingURL=docx-parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"docx-parser.d.ts","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAU1D,MAAM,WAAW,cAAc;IAC7B,OAAO,EAAE,MAAM,CAAA;IACf,YAAY,EAAE,SAAS,eAAe,EAAE,CAAA;IACxC,YAAY,CAAC,EAAE,SAAS;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,EAAE,CAAA;CACjE;AAqQD,wBAAsB,oBAAoB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,SAAS,CAAC,CAkBtF;AAED,wBAAgB,yBAAyB,CAAC,QAAQ,EAAE,QAAQ,GAAG,cAAc,CAgD5E"}
|
|
@@ -0,0 +1,328 @@
|
|
|
1
|
+
import { JSDOM } from 'jsdom';
|
|
2
|
+
import JSZip from 'jszip';
|
|
3
|
+
const CORE_TITLE_NAMESPACE = 'http://purl.org/dc/elements/1.1/';
|
|
4
|
+
const PROSE_BLOCK_TAGS = new Set(['H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'P', 'BLOCKQUOTE', 'PRE']);
|
|
5
|
+
const LIST_TAGS = new Set(['UL', 'OL']);
|
|
6
|
+
const SKIPPED_TAGS = new Set(['SCRIPT', 'STYLE', 'NOSCRIPT']);
|
|
7
|
+
const IMAGE_MARKER_PREFIX = '\u0000rag-image:';
|
|
8
|
+
const IMAGE_MARKER_SUFFIX = '\u0000';
|
|
9
|
+
const IMAGE_MARKER_PATTERN = new RegExp(`${IMAGE_MARKER_PREFIX}(\\d+)${IMAGE_MARKER_SUFFIX}`, 'gu');
|
|
10
|
+
function imageMarker(element) {
|
|
11
|
+
if (element.tagName !== 'IMG')
|
|
12
|
+
return '';
|
|
13
|
+
const value = element.getAttribute('data-rag-image-index');
|
|
14
|
+
return value !== null && /^\d+$/.test(value)
|
|
15
|
+
? `${IMAGE_MARKER_PREFIX}${value}${IMAGE_MARKER_SUFFIX}`
|
|
16
|
+
: '';
|
|
17
|
+
}
|
|
18
|
+
function normalizeSingleLine(text) {
|
|
19
|
+
return text.replace(/\s+/g, ' ').trim();
|
|
20
|
+
}
|
|
21
|
+
function normalizeTextContent(element) {
|
|
22
|
+
const parts = [];
|
|
23
|
+
const visit = (node) => {
|
|
24
|
+
if (node.nodeType === 3) {
|
|
25
|
+
parts.push(node.nodeValue ?? '');
|
|
26
|
+
return;
|
|
27
|
+
}
|
|
28
|
+
if (node.nodeType !== 1)
|
|
29
|
+
return;
|
|
30
|
+
const child = node;
|
|
31
|
+
if (SKIPPED_TAGS.has(child.tagName))
|
|
32
|
+
return;
|
|
33
|
+
const marker = imageMarker(child);
|
|
34
|
+
if (marker) {
|
|
35
|
+
parts.push(marker);
|
|
36
|
+
return;
|
|
37
|
+
}
|
|
38
|
+
if (child.tagName === 'BR') {
|
|
39
|
+
parts.push(' ');
|
|
40
|
+
return;
|
|
41
|
+
}
|
|
42
|
+
const ownsTextBoundary = PROSE_BLOCK_TAGS.has(child.tagName) || LIST_TAGS.has(child.tagName) || child.tagName === 'LI';
|
|
43
|
+
if (ownsTextBoundary)
|
|
44
|
+
parts.push(' ');
|
|
45
|
+
for (const descendant of child.childNodes)
|
|
46
|
+
visit(descendant);
|
|
47
|
+
if (ownsTextBoundary)
|
|
48
|
+
parts.push(' ');
|
|
49
|
+
};
|
|
50
|
+
for (const child of element.childNodes)
|
|
51
|
+
visit(child);
|
|
52
|
+
return normalizeSingleLine(parts.join(''));
|
|
53
|
+
}
|
|
54
|
+
function serializeInlineText(element) {
|
|
55
|
+
const parts = [];
|
|
56
|
+
const visit = (node) => {
|
|
57
|
+
if (node.nodeType === 3) {
|
|
58
|
+
parts.push(node.nodeValue ?? '');
|
|
59
|
+
return;
|
|
60
|
+
}
|
|
61
|
+
if (node.nodeType !== 1)
|
|
62
|
+
return;
|
|
63
|
+
const child = node;
|
|
64
|
+
if (SKIPPED_TAGS.has(child.tagName))
|
|
65
|
+
return;
|
|
66
|
+
const marker = imageMarker(child);
|
|
67
|
+
if (marker) {
|
|
68
|
+
parts.push(marker);
|
|
69
|
+
return;
|
|
70
|
+
}
|
|
71
|
+
if (child.tagName === 'BR') {
|
|
72
|
+
parts.push('\n');
|
|
73
|
+
return;
|
|
74
|
+
}
|
|
75
|
+
for (const descendant of child.childNodes)
|
|
76
|
+
visit(descendant);
|
|
77
|
+
};
|
|
78
|
+
for (const child of element.childNodes)
|
|
79
|
+
visit(child);
|
|
80
|
+
const lines = parts.join('').split('\n').map(normalizeSingleLine);
|
|
81
|
+
while (lines[0] === '')
|
|
82
|
+
lines.shift();
|
|
83
|
+
while (lines.at(-1) === '')
|
|
84
|
+
lines.pop();
|
|
85
|
+
return lines.join('\n');
|
|
86
|
+
}
|
|
87
|
+
function serializeList(list) {
|
|
88
|
+
const lines = [];
|
|
89
|
+
const serializeListElement = (currentList) => {
|
|
90
|
+
for (const child of currentList.children) {
|
|
91
|
+
if (child.tagName !== 'LI')
|
|
92
|
+
continue;
|
|
93
|
+
let currentLine = '';
|
|
94
|
+
const flush = () => {
|
|
95
|
+
const normalized = normalizeSingleLine(currentLine);
|
|
96
|
+
if (normalized)
|
|
97
|
+
lines.push(normalized);
|
|
98
|
+
currentLine = '';
|
|
99
|
+
};
|
|
100
|
+
const visitItemNode = (node) => {
|
|
101
|
+
if (node.nodeType === 3) {
|
|
102
|
+
currentLine += node.nodeValue ?? '';
|
|
103
|
+
return;
|
|
104
|
+
}
|
|
105
|
+
if (node.nodeType !== 1)
|
|
106
|
+
return;
|
|
107
|
+
const element = node;
|
|
108
|
+
if (SKIPPED_TAGS.has(element.tagName))
|
|
109
|
+
return;
|
|
110
|
+
const marker = imageMarker(element);
|
|
111
|
+
if (marker) {
|
|
112
|
+
currentLine += marker;
|
|
113
|
+
return;
|
|
114
|
+
}
|
|
115
|
+
if (element.tagName === 'BR') {
|
|
116
|
+
flush();
|
|
117
|
+
return;
|
|
118
|
+
}
|
|
119
|
+
if (LIST_TAGS.has(element.tagName)) {
|
|
120
|
+
flush();
|
|
121
|
+
serializeListElement(element);
|
|
122
|
+
return;
|
|
123
|
+
}
|
|
124
|
+
for (const descendant of element.childNodes)
|
|
125
|
+
visitItemNode(descendant);
|
|
126
|
+
};
|
|
127
|
+
for (const itemChild of child.childNodes)
|
|
128
|
+
visitItemNode(itemChild);
|
|
129
|
+
flush();
|
|
130
|
+
}
|
|
131
|
+
};
|
|
132
|
+
serializeListElement(list);
|
|
133
|
+
return lines.join('\n');
|
|
134
|
+
}
|
|
135
|
+
function directTableRows(table) {
|
|
136
|
+
return Array.from(table.querySelectorAll('tr')).filter((row) => row.closest('table') === table);
|
|
137
|
+
}
|
|
138
|
+
function directRowCells(row) {
|
|
139
|
+
return Array.from(row.children).filter((cell) => cell.tagName === 'TH' || cell.tagName === 'TD');
|
|
140
|
+
}
|
|
141
|
+
function hasSpanningCell(cells) {
|
|
142
|
+
return cells.some((cell) => {
|
|
143
|
+
const rowSpan = Number.parseInt(cell.getAttribute('rowspan') ?? '1', 10);
|
|
144
|
+
const columnSpan = Number.parseInt(cell.getAttribute('colspan') ?? '1', 10);
|
|
145
|
+
return rowSpan > 1 || columnSpan > 1;
|
|
146
|
+
});
|
|
147
|
+
}
|
|
148
|
+
function isSupportedTable(table, rows) {
|
|
149
|
+
if (table.querySelector('table') !== null || rows.length < 2)
|
|
150
|
+
return false;
|
|
151
|
+
const rowCells = rows.map(directRowCells);
|
|
152
|
+
const columnCount = rowCells[0]?.length ?? 0;
|
|
153
|
+
return (columnCount > 0 &&
|
|
154
|
+
rowCells.every((cells) => cells.length === columnCount && !hasSpanningCell(cells)));
|
|
155
|
+
}
|
|
156
|
+
function collectTextTokens(element) {
|
|
157
|
+
const tokens = [];
|
|
158
|
+
const visit = (node) => {
|
|
159
|
+
if (node.nodeType === 3) {
|
|
160
|
+
const token = normalizeSingleLine(node.nodeValue ?? '');
|
|
161
|
+
if (token)
|
|
162
|
+
tokens.push(token);
|
|
163
|
+
return;
|
|
164
|
+
}
|
|
165
|
+
if (node.nodeType !== 1)
|
|
166
|
+
return;
|
|
167
|
+
const child = node;
|
|
168
|
+
if (SKIPPED_TAGS.has(child.tagName))
|
|
169
|
+
return;
|
|
170
|
+
const marker = imageMarker(child);
|
|
171
|
+
if (marker) {
|
|
172
|
+
tokens.push(marker);
|
|
173
|
+
return;
|
|
174
|
+
}
|
|
175
|
+
for (const descendant of child.childNodes)
|
|
176
|
+
visit(descendant);
|
|
177
|
+
};
|
|
178
|
+
visit(element);
|
|
179
|
+
return tokens.join(' ');
|
|
180
|
+
}
|
|
181
|
+
function emitTable(table) {
|
|
182
|
+
const rows = directTableRows(table);
|
|
183
|
+
const headerRow = rows[0];
|
|
184
|
+
if (!isSupportedTable(table, rows) || headerRow === undefined) {
|
|
185
|
+
const text = collectTextTokens(table);
|
|
186
|
+
return text ? [{ text, atomic: false }] : [];
|
|
187
|
+
}
|
|
188
|
+
// Ordinary Word tables often contain only td cells, so the approved contract
|
|
189
|
+
// deliberately treats the first physical row as labels even without th markup.
|
|
190
|
+
const headerMarkers = [];
|
|
191
|
+
const headers = directRowCells(headerRow).map((cell, index) => {
|
|
192
|
+
const text = normalizeTextContent(cell);
|
|
193
|
+
headerMarkers.push(...(text.match(IMAGE_MARKER_PATTERN) ?? []));
|
|
194
|
+
return text.replace(IMAGE_MARKER_PATTERN, '') || `Column ${index + 1}`;
|
|
195
|
+
});
|
|
196
|
+
const rowBlocks = rows.slice(1).flatMap((row) => {
|
|
197
|
+
const values = directRowCells(row);
|
|
198
|
+
const text = values
|
|
199
|
+
.map((cell, index) => `${headers[index] ?? `Column ${index + 1}`}: ${normalizeTextContent(cell)}`)
|
|
200
|
+
.join('\n');
|
|
201
|
+
return text ? [{ text, atomic: true }] : [];
|
|
202
|
+
});
|
|
203
|
+
const first = rowBlocks[0];
|
|
204
|
+
if (first && headerMarkers.length > 0)
|
|
205
|
+
first.text = headerMarkers.join('') + first.text;
|
|
206
|
+
return rowBlocks;
|
|
207
|
+
}
|
|
208
|
+
function hasRecognizedBlockDescendant(element) {
|
|
209
|
+
return Array.from(element.querySelectorAll('*')).some((descendant) => PROSE_BLOCK_TAGS.has(descendant.tagName) ||
|
|
210
|
+
LIST_TAGS.has(descendant.tagName) ||
|
|
211
|
+
descendant.tagName === 'TABLE');
|
|
212
|
+
}
|
|
213
|
+
function emitDocumentBlocks(document) {
|
|
214
|
+
const blocks = [];
|
|
215
|
+
const emitElement = (element) => {
|
|
216
|
+
if (SKIPPED_TAGS.has(element.tagName))
|
|
217
|
+
return;
|
|
218
|
+
if (PROSE_BLOCK_TAGS.has(element.tagName)) {
|
|
219
|
+
const text = serializeInlineText(element);
|
|
220
|
+
if (text)
|
|
221
|
+
blocks.push({ text, atomic: false });
|
|
222
|
+
return;
|
|
223
|
+
}
|
|
224
|
+
if (LIST_TAGS.has(element.tagName)) {
|
|
225
|
+
const text = serializeList(element);
|
|
226
|
+
if (text)
|
|
227
|
+
blocks.push({ text, atomic: false });
|
|
228
|
+
return;
|
|
229
|
+
}
|
|
230
|
+
if (element.tagName === 'TABLE') {
|
|
231
|
+
blocks.push(...emitTable(element));
|
|
232
|
+
return;
|
|
233
|
+
}
|
|
234
|
+
if (!hasRecognizedBlockDescendant(element)) {
|
|
235
|
+
const text = normalizeTextContent(element);
|
|
236
|
+
if (text)
|
|
237
|
+
blocks.push({ text, atomic: false });
|
|
238
|
+
return;
|
|
239
|
+
}
|
|
240
|
+
for (const child of element.childNodes)
|
|
241
|
+
emitNode(child);
|
|
242
|
+
};
|
|
243
|
+
const emitNode = (node) => {
|
|
244
|
+
if (node.nodeType === 3) {
|
|
245
|
+
const text = normalizeSingleLine(node.nodeValue ?? '');
|
|
246
|
+
if (text)
|
|
247
|
+
blocks.push({ text, atomic: false });
|
|
248
|
+
return;
|
|
249
|
+
}
|
|
250
|
+
if (node.nodeType === 1)
|
|
251
|
+
emitElement(node);
|
|
252
|
+
};
|
|
253
|
+
for (const node of document.body.childNodes)
|
|
254
|
+
emitNode(node);
|
|
255
|
+
return blocks;
|
|
256
|
+
}
|
|
257
|
+
export async function extractDocxCoreTitle(buffer) {
|
|
258
|
+
const zip = await JSZip.loadAsync(buffer);
|
|
259
|
+
const coreProperties = zip.file('docProps/core.xml');
|
|
260
|
+
if (!coreProperties)
|
|
261
|
+
return undefined;
|
|
262
|
+
const xml = await coreProperties.async('string');
|
|
263
|
+
try {
|
|
264
|
+
const document = new JSDOM(xml, { contentType: 'text/xml' }).window.document;
|
|
265
|
+
if (document.querySelector('parsererror'))
|
|
266
|
+
return undefined;
|
|
267
|
+
const title = document.getElementsByTagNameNS(CORE_TITLE_NAMESPACE, 'title').item(0);
|
|
268
|
+
const normalized = normalizeSingleLine(title?.textContent ?? '');
|
|
269
|
+
return normalized || undefined;
|
|
270
|
+
}
|
|
271
|
+
catch {
|
|
272
|
+
// Core properties are optional display metadata; malformed XML falls
|
|
273
|
+
// through to the heading/filename chain without rejecting readable body text.
|
|
274
|
+
return undefined;
|
|
275
|
+
}
|
|
276
|
+
}
|
|
277
|
+
export function convertDocxDocumentToText(document) {
|
|
278
|
+
const blocks = emitDocumentBlocks(document);
|
|
279
|
+
let content = '';
|
|
280
|
+
const atomicRanges = [];
|
|
281
|
+
const imageAnchors = [];
|
|
282
|
+
for (const block of blocks) {
|
|
283
|
+
const matches = [...block.text.matchAll(IMAGE_MARKER_PATTERN)];
|
|
284
|
+
if (matches.length === 0) {
|
|
285
|
+
if (content)
|
|
286
|
+
content += '\n\n';
|
|
287
|
+
const start = content.length;
|
|
288
|
+
content += block.text;
|
|
289
|
+
if (block.atomic)
|
|
290
|
+
atomicRanges.push({ start, end: content.length });
|
|
291
|
+
continue;
|
|
292
|
+
}
|
|
293
|
+
const cleanText = block.text.replace(IMAGE_MARKER_PATTERN, '');
|
|
294
|
+
const normalizedText = cleanText.split('\n').map(normalizeSingleLine).join('\n');
|
|
295
|
+
if (!normalizedText) {
|
|
296
|
+
for (const match of matches) {
|
|
297
|
+
imageAnchors.push({ offset: content.length, imageIndex: Number(match[1]) });
|
|
298
|
+
}
|
|
299
|
+
continue;
|
|
300
|
+
}
|
|
301
|
+
if (content)
|
|
302
|
+
content += '\n\n';
|
|
303
|
+
const start = content.length;
|
|
304
|
+
for (const match of matches) {
|
|
305
|
+
const rawOffset = match.index ?? 0;
|
|
306
|
+
const cleanPrefix = block.text.slice(0, rawOffset).replace(IMAGE_MARKER_PATTERN, '');
|
|
307
|
+
const normalizedPrefix = cleanPrefix.split('\n').map(normalizeSingleLine).join('\n');
|
|
308
|
+
imageAnchors.push({ offset: start + normalizedPrefix.length, imageIndex: Number(match[1]) });
|
|
309
|
+
}
|
|
310
|
+
content += normalizedText;
|
|
311
|
+
if (block.atomic)
|
|
312
|
+
atomicRanges.push({ start, end: content.length });
|
|
313
|
+
}
|
|
314
|
+
imageAnchors.sort((left, right) => left.imageIndex - right.imageIndex);
|
|
315
|
+
const seenImageIndices = new Set();
|
|
316
|
+
const uniqueImageAnchors = imageAnchors.filter((anchor) => {
|
|
317
|
+
if (seenImageIndices.has(anchor.imageIndex))
|
|
318
|
+
return false;
|
|
319
|
+
seenImageIndices.add(anchor.imageIndex);
|
|
320
|
+
return true;
|
|
321
|
+
});
|
|
322
|
+
return {
|
|
323
|
+
content,
|
|
324
|
+
atomicRanges,
|
|
325
|
+
...(uniqueImageAnchors.length === 0 ? {} : { imageAnchors: uniqueImageAnchors }),
|
|
326
|
+
};
|
|
327
|
+
}
|
|
328
|
+
//# sourceMappingURL=docx-parser.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"docx-parser.js","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,KAAK,MAAM,OAAO,CAAA;AAGzB,MAAM,oBAAoB,GAAG,kCAAkC,CAAA;AAC/D,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,YAAY,EAAE,KAAK,CAAC,CAAC,CAAA;AAChG,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;AACvC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC,CAAA;AAC7D,MAAM,mBAAmB,GAAG,kBAAkB,CAAA;AAC9C,MAAM,mBAAmB,GAAG,QAAQ,CAAA;AACpC,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,mBAAmB,SAAS,mBAAmB,EAAE,EAAE,IAAI,CAAC,CAAA;AAanG,SAAS,WAAW,CAAC,OAAgB;IACnC,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK;QAAE,OAAO,EAAE,CAAA;IACxC,MAAM,KAAK,GAAG,OAAO,CAAC,YAAY,CAAC,sBAAsB,CAAC,CAAA;IAC1D,OAAO,KAAK,KAAK,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC;QAC1C,CAAC,CAAC,GAAG,mBAAmB,GAAG,KAAK,GAAG,mBAAmB,EAAE;QACxD,CAAC,CAAC,EAAE,CAAA;AACR,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACzC,CAAC;AAED,SAAS,oBAAoB,CAAC,OAAgB;IAC5C,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YAChC,OAAM;QACR,CAAC;QACD,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;YAAE,OAAM;QAE/B,MAAM,KAAK,GAAG,IAAe,CAAA;QAC7B,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAAE,OAAM;QAC3C,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAClB,OAAM;QACR,CAAC;QACD,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACf,OAAM;QACR,CAAC;QAED,MAAM,gBAAgB,GACpB,gBAAgB,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAA;QAC/F,IAAI,gBAAgB;YAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACrC,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU;YAAE,KAAK,CAAC,UAAU,CAAC,CAAA;QAC5D,IAAI,gBAAgB;YAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACvC,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU;QAAE,KAAK,CAAC,KAAK,CAAC,CAAA;IACpD,OAAO,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;AAC5C,CAAC;AAED,SAAS,mBAAmB,CAAC,OAAgB;IAC3C,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YAChC,OAAM;QACR,CAAC;QACD,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;YAAE,OAAM;QAE/B,MAAM,KAAK,GAAG,IAAe,CAAA;QAC7B,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAAE,OAAM;QAC3C,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAClB,OAAM;QACR,CAAC;QACD,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAChB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU;YAAE,KAAK,CAAC,UAAU,CAAC,CAAA;IAC9D,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU;QAAE,KAAK,CAAC,KAAK,CAAC,CAAA;IAEpD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IACjE,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE;QAAE,KAAK,CAAC,KAAK,EAAE,CAAA;IACrC,OAAO,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE;QAAE,KAAK,CAAC,GAAG,EAAE,CAAA;IACvC,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAClC,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,oBAAoB,GAAG,CAAC,WAAoB,EAAQ,EAAE;QAC1D,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,QAAQ,EAAE,CAAC;YACzC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI;gBAAE,SAAQ;YAEpC,IAAI,WAAW,GAAG,EAAE,CAAA;YACpB,MAAM,KAAK,GAAG,GAAS,EAAE;gBACvB,MAAM,UAAU,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAA;gBACnD,IAAI,UAAU;oBAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;gBACtC,WAAW,GAAG,EAAE,CAAA;YAClB,CAAC,CAAA;YACD,MAAM,aAAa,GAAG,CAAC,IAAU,EAAQ,EAAE;gBACzC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;oBACxB,WAAW,IAAI,IAAI,CAAC,SAAS,IAAI,EAAE,CAAA;oBACnC,OAAM;gBACR,CAAC;gBACD,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;oBAAE,OAAM;gBAE/B,MAAM,OAAO,GAAG,IAAe,CAAA;gBAC/B,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC;oBAAE,OAAM;gBAC7C,MAAM,MAAM,GAAG,WAAW,CAAC,OAAO,CAAC,CAAA;gBACnC,IAAI,MAAM,EAAE,CAAC;oBACX,WAAW,IAAI,MAAM,CAAA;oBACrB,OAAM;gBACR,CAAC;gBACD,IAAI,OAAO,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;oBAC7B,KAAK,EAAE,CAAA;oBACP,OAAM;gBACR,CAAC;gBACD,IAAI,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;oBACnC,KAAK,EAAE,CAAA;oBACP,oBAAoB,CAAC,OAAO,CAAC,CAAA;oBAC7B,OAAM;gBACR,CAAC;gBACD,KAAK,MAAM,UAAU,IAAI,OAAO,CAAC,UAAU;oBAAE,aAAa,CAAC,UAAU,CAAC,CAAA;YACxE,CAAC,CAAA;YAED,KAAK,MAAM,SAAS,IAAI,KAAK,CAAC,UAAU;gBAAE,aAAa,CAAC,SAAS,CAAC,CAAA;YAClE,KAAK,EAAE,CAAA;QACT,CAAC;IACH,CAAC,CAAA;IAED,oBAAoB,CAAC,IAAI,CAAC,CAAA;IAC1B,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAA;AACjG,CAAC;AAED,SAAS,cAAc,CAAC,GAAY;IAClC,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;AAClG,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB;IAChD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QACxE,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QAC3E,OAAO,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,CAAA;IACtC,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAc,EAAE,IAAwB;IAChE,IAAI,KAAK,CAAC,aAAa,CAAC,OAAO,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,KAAK,CAAA;IAE1E,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IACzC,MAAM,WAAW,GAAG,QAAQ,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;IAC5C,OAAO,CACL,WAAW,GAAG,CAAC;QACf,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,WAAW,IAAI,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC,CACnF,CAAA;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAgB;IACzC,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YACvD,IAAI,KAAK;gBAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YAC7B,OAAM;QACR,CAAC;QACD,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;YAAE,OAAM;QAE/B,MAAM,KAAK,GAAG,IAAe,CAAA;QAC7B,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAAE,OAAM;QAC3C,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACnB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU;YAAE,KAAK,CAAC,UAAU,CAAC,CAAA;IAC9D,CAAC,CAAA;IACD,KAAK,CAAC,OAAO,CAAC,CAAA;IACd,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,SAAS,CAAC,KAAc;IAC/B,MAAM,IAAI,GAAG,eAAe,CAAC,KAAK,CAAC,CAAA;IACnC,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACzB,IAAI,CAAC,gBAAgB,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC9D,MAAM,IAAI,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAA;QACrC,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC9C,CAAC;IAED,6EAA6E;IAC7E,+EAA+E;IAC/E,MAAM,aAAa,GAAa,EAAE,CAAA;IAClC,MAAM,OAAO,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5D,MAAM,IAAI,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAA;QACvC,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;QAC/D,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAA;IACxE,CAAC,CAAC,CAAA;IAEF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC9C,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;QAClC,MAAM,IAAI,GAAG,MAAM;aAChB,GAAG,CACF,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,KAAK,oBAAoB,CAAC,IAAI,CAAC,EAAE,CAC7F;aACA,IAAI,CAAC,IAAI,CAAC,CAAA;QACb,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC7C,CAAC,CAAC,CAAA;IACF,MAAM,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAA;IAC1B,IAAI,KAAK,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC;QAAE,KAAK,CAAC,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAA;IACvF,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,SAAS,4BAA4B,CAAC,OAAgB;IACpD,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CACnD,CAAC,UAAU,EAAE,EAAE,CACb,gBAAgB,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACxC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,UAAU,CAAC,OAAO,KAAK,OAAO,CACjC,CAAA;AACH,CAAC;AAED,SAAS,kBAAkB,CAAC,QAAkB;IAC5C,MAAM,MAAM,GAAmB,EAAE,CAAA;IAEjC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAQ,EAAE;QAC7C,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC;YAAE,OAAM;QAC7C,IAAI,gBAAgB,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1C,MAAM,IAAI,GAAG,mBAAmB,CAAC,OAAO,CAAC,CAAA;YACzC,IAAI,IAAI;gBAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;YAC9C,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YACnC,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,CAAC,CAAA;YACnC,IAAI,IAAI;gBAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;YAC9C,OAAM;QACR,CAAC;QACD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAClC,OAAM;QACR,CAAC;QACD,IAAI,CAAC,4BAA4B,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3C,MAAM,IAAI,GAAG,oBAAoB,CAAC,OAAO,CAAC,CAAA;YAC1C,IAAI,IAAI;gBAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;YAC9C,OAAM;QACR,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU;YAAE,QAAQ,CAAC,KAAK,CAAC,CAAA;IACzD,CAAC,CAAA;IAED,MAAM,QAAQ,GAAG,CAAC,IAAU,EAAQ,EAAE;QACpC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,GAAG,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YACtD,IAAI,IAAI;gBAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;YAC9C,OAAM;QACR,CAAC;QACD,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;YAAE,WAAW,CAAC,IAAe,CAAC,CAAA;IACvD,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,IAAI,CAAC,UAAU;QAAE,QAAQ,CAAC,IAAI,CAAC,CAAA;IAC3D,OAAO,MAAM,CAAA;AACf,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,MAAc;IACvD,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,cAAc,GAAG,GAAG,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAA;IACpD,IAAI,CAAC,cAAc;QAAE,OAAO,SAAS,CAAA;IAErC,MAAM,GAAG,GAAG,MAAM,cAAc,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAChD,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,IAAI,KAAK,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;QAC5E,IAAI,QAAQ,CAAC,aAAa,CAAC,aAAa,CAAC;YAAE,OAAO,SAAS,CAAA;QAE3D,MAAM,KAAK,GAAG,QAAQ,CAAC,sBAAsB,CAAC,oBAAoB,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACpF,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,EAAE,WAAW,IAAI,EAAE,CAAC,CAAA;QAChE,OAAO,UAAU,IAAI,SAAS,CAAA;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,qEAAqE;QACrE,8EAA8E;QAC9E,OAAO,SAAS,CAAA;IAClB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,yBAAyB,CAAC,QAAkB;IAC1D,MAAM,MAAM,GAAG,kBAAkB,CAAC,QAAQ,CAAC,CAAA;IAC3C,IAAI,OAAO,GAAG,EAAE,CAAA;IAChB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,YAAY,GAA6C,EAAE,CAAA;IAEjE,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAA;QAC9D,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACzB,IAAI,OAAO;gBAAE,OAAO,IAAI,MAAM,CAAA;YAC9B,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;YAC5B,OAAO,IAAI,KAAK,CAAC,IAAI,CAAA;YACrB,IAAI,KAAK,CAAC,MAAM;gBAAE,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;YACnE,SAAQ;QACV,CAAC;QACD,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,CAAA;QAC9D,MAAM,cAAc,GAAG,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAChF,IAAI,CAAC,cAAc,EAAE,CAAC;YACpB,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;gBAC5B,YAAY,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;YAC7E,CAAC;YACD,SAAQ;QACV,CAAC;QAED,IAAI,OAAO;YAAE,OAAO,IAAI,MAAM,CAAA;QAC9B,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;QAC5B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,IAAI,CAAC,CAAA;YAClC,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,CAAA;YACpF,MAAM,gBAAgB,GAAG,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACpF,YAAY,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,GAAG,gBAAgB,CAAC,MAAM,EAAE,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;QAC9F,CAAC;QACD,OAAO,IAAI,cAAc,CAAA;QACzB,IAAI,KAAK,CAAC,MAAM;YAAE,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IACrE,CAAC;IAED,YAAY,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC,CAAA;IACtE,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAU,CAAA;IAC1C,MAAM,kBAAkB,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;QACxD,IAAI,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC;YAAE,OAAO,KAAK,CAAA;QACzD,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;QACvC,OAAO,IAAI,CAAA;IACb,CAAC,CAAC,CAAA;IACF,OAAO;QACL,OAAO;QACP,YAAY;QACZ,GAAG,CAAC,kBAAkB,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,kBAAkB,EAAE,CAAC;KACjF,CAAA;AACH,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Parse HTML content and extract main content as Markdown
|
|
3
|
+
*
|
|
4
|
+
* Flow:
|
|
5
|
+
* 1. HTML string → JSDOM (DOM creation)
|
|
6
|
+
* 2. JSDOM → Readability (main content extraction, noise removal)
|
|
7
|
+
* 3. Readability result → Turndown (Markdown conversion)
|
|
8
|
+
* 4. Title extracted separately via extractHtmlTitle (NOT prepended to content)
|
|
9
|
+
*
|
|
10
|
+
* @param html - Raw HTML string
|
|
11
|
+
* @param url - Source URL (used for resolving relative links)
|
|
12
|
+
* @returns Object with content (markdown) and title (extracted separately)
|
|
13
|
+
*/
|
|
14
|
+
export declare function parseHtml(html: string, url: string): Promise<{
|
|
15
|
+
content: string;
|
|
16
|
+
title: string;
|
|
17
|
+
}>;
|
|
18
|
+
//# sourceMappingURL=html-parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"html-parser.d.ts","sourceRoot":"","sources":["../../src/parser/html-parser.ts"],"names":[],"mappings":"AAuDA;;;;;;;;;;;;GAYG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,MAAM,EACZ,GAAG,EAAE,MAAM,GACV,OAAO,CAAC;IAAE,OAAO,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CA0D7C"}
|